
What strike rate leaves out
Strike rate is the share of results classified as wins under a stated definition. It does not measure profit by itself. To interpret a football tipster’s record, you also need the stakes, prices, settlement rules and period covered.
Specify how refunds, voids, half-wins and half-losses are handled. Two summaries can show different win rates for the same set of entries if they use different denominators.
A hypothetical high-win-rate loss
Suppose 100 fully settled bets each stake one unit, every bet has decimal odds of 1.30, and exactly 70 win. The winning returns total 70 × 1.30 = 91 units, including their stakes. Total stakes are 100 units, so net profit is -9 units and realised ROI is -9%.
The same arithmetic with exactly 45 wins from 100 one-unit bets at 2.50 produces 112.5 units returned, 12.5 units net profit and +12.5% realised ROI. These are constructed examples, not OptaTips performance data or forecasts. Decimal returns include the original stake; see Betfair’s decimal-odds explanation.
Separate realised ROI from expected return
Realised ROI is net profit divided by total stakes, multiplied by 100%. It describes the completed sample. If a bet has only win-or-lose outcomes, decimal odds d and a genuine win probability p, its expected net return per unit is p × d – 1.
An observed win rate is an estimate from a sample, not a known future probability. Substituting it into an expected-return formula does not prove that the next bets have a positive expectation. Prices and selection methods may also change.
Why averages need care
When odds or stakes vary, calculate profit from each settled bet before adding the totals. Multiplying an overall win rate by the average of all quoted odds can give the wrong realised ROI because the winning bets may have different prices from the losing bets.
For a hypothetical two-bet example, a one-unit winner at 1.50 and a one-unit loser at 3.50 return only 1.50 units from 2 staked: ROI is -25%, despite a 50% win rate and average quoted odds of 2.50.
Read the whole record
Compare equal periods and explain costs, refunds and staking conventions. Include losing runs and unresolved selections rather than selecting a favourable month. No fixed sample size proves that a profitable period will persist.
Use the profile metrics checklist to examine the other labels beside win rate. A historical percentage is evidence about its stated sample, not a promise of future profit.
Comments 0
No comments yet.